Social Media Sentiment Analysis using Deep Learning: Review

The paper explores the application of deep learning techniques to sentiment analysis on social media platforms. It contrasts these advanced methods with conventional machine learning algorithms and highlights the advantages of using deep neural networks to classify more nuanced and accurate emotions. The study investigates the accuracy of different deep learning models, including RNN, CNN, LSTM, and transform-based models such as BERT and GPT, in processing complex and unstructured social media data. It also discusses the importance of preprocessing to improve model performance and presents a comparative analysis of accuracy of different models through a literature study. The document concludes by identifying research gaps, determining the most accurate sentiment analysis model, and suggesting directions for future methodological improvements.